09 Wilhelmus Final 11/9/01 11:17 AM Page 111 CURVULARIA KERATITIS* BY Kirk R. Wilhelmus, MD, MPH, AND Dan B. Jones, MD ABSTRACT Purpose: To determine the risk factors and clinical signs of Curvularia keratitis and to evaluate the management and out- come of this corneal phæohyphomycosis. Methods: We reviewed clinical and laboratory records from 1970 to 1999 to identify patients treated at our institution for culture-proven Curvularia keratitis. Descriptive statistics and regression models were used to identify variables associ- ated with the length of antifungal therapy and with visual outcome. In vitro susceptibilities were compared to the clini- cal results obtained with topical natamycin. Results: During the 30-year period, our laboratory isolated and identified Curvularia from 43 patients with keratitis, of whom 32 individuals were treated and followed up at our institute and whose data were analyzed. Trauma, usually with plants or dirt, was the risk factor in one half; and 69% occurred during the hot, humid summer months along the US Gulf Coast. Presenting signs varied from superficial, feathery infiltrates of the central cornea to suppurative ulceration of the peripheral cornea. A hypopyon was unusual, occurring in only 4 (12%) of the eyes but indicated a significantly (P = .01) increased risk of subsequent complications. The sensitivity of stained smears of corneal scrapings was 78%. Curvularia could be detected by a panfungal polymerase chain reaction. Fungi were detected on blood or chocolate agar at or before the time that growth occurred on Sabouraud agar or in brain-heart infusion in 83% of cases, although colonies appeared only on the fungal media from the remaining 4 sets of specimens. Curvularia was the third most prevalent filamentous fungus among our corneal isolates and the most common dematiaceous mold. Corneal isolates included C senegalensis, C lunata, C pallescens, and C prasadii. All tested isolates were inhibited by 4 ␮g/mL or less of natamycin. Topical natamycin was used for a median duration of 1 month, but a delay in diagnosis beyond 1 week dou- bled the average length of topical antifungal treatment (P =.005). Visual acuity improved to 20/40 or better in 25 (78%) of the eyes. Conclusions: Curvularia keratitis typically presented as superficial feathery infiltration, rarely with visible pigmentation, that gradually became focally suppurative. Smears of corneal scrapings often disclosed hyphae, and culture media showed dematiaceous fungal growth within 1 week. Natamycin had excellent in vitro activity and led to clinical resolu- tion with good vision in most patients with corneal curvulariosis. Complications requiring surgery were not common but included exophytic inflammatory fungal sequestration, treated by superficial lamellar keratectomy, and corneal perfora- tion, managed by penetrating keratoplasty. Tr Am Ophth Soc 2001;99:111-132 INTRODUCTION Curvularia is a prevalent member of these darkly pig- mented fungi that received its current name in 19338 and Fungal infections of the eye are a growing threat that have that is related to the sexual teleomorph Cochliobolus. substantial morbidity and cost.1 Aspergillus and Fusarium This genus of filamentous fungi colonizes soil and are long recognized as ocular pathogens,2 but the demati- vegetation and spreads by airborne spores. Some of the 40 aceous hyphomycetes have emerged as important oppor- Curvularia species are phytopathogens. Plant diseases tunists.3-6 Originally named for their tufted, floccose range from seedling failure to leaf blight,9 including grass appearance in culture, dematiaceous fungi comprise those “fade out” during hot, humid weather. Curvularial growth septate molds with melanin in their hyphae and conidia.7 on stored grain, thatch, and other dead plant material looks like smudges of blackish dust. *From the Sid W. Richardson Ocular Microbiology Laboratory, Several Curvularia species are zoopathogenic. Department of Ophthalmology, Cullen Eye Institute, Baylor College of Wound infection is the most common disease caused by Medicine, Houston, Texas. Supported by a clinical investigator award Curvularia and ranges from onychomycosis to skin ulcer- (EY00377), cooperative agreement (EY09696), and core grant ation and subcutaneous mycetoma.10,11 Other human (EY02520) from the National Eye Institute; a senior scientific investiga- tor award from the Research to Prevent Blindness, Inc; and the Sid Curvularia infections are invasive and allergic sinusitis Richardson Foundation. and bronchopulmonary disease. Abscesses of the lung, Tr. Am. Ophth. Soc. Vol. 99, 2001 111 09 Wilhelmus Final 11/9/01 11:17 AM Page 112 Wilhelmus et al brain, liver, and connective tissue have occurred. 1B). Rapidly growing mycelia often produced a central Nosocomial infections include dialysis-related peritonitis depression in the dark, matted colony. Slide culture and postsurgical endocarditis.12 showed the characteristic microscopic appearance of Infection of the cornea, reported in 1959,13 was the branched, septate, tawny hyphae and short, nodose, first human disease proved to be caused by Curvularia. brown conidiophores bearing single and clustered septate Other ocular infections consist of conjunctivitis,14 dacry- conidia (Fig 1C). Speciation was based on the microscop- ocystitis,15,16 sino-orbital cellulitis,17 and endophthalmi- ic appearance of conidia.90,91 The minimum inhibitory con- tis.14,18-20 But the cornea is the most commonly infected centration (MIC) was determined for selected isolates site.2,3,13,14,21-88 To describe the clinical spectrum and man- with antibiotic-saturated paper discs in multiwell plates92 agement of Curvularia keratitis, we reviewed our experi- ence with this corneal phæohyphomycosis. METHODS Cases of culture-positive Curvularia keratitis were identi- fied by reviewing the records of our ocular microbiology laboratory for patients with keratomycosis. Patients evalu- ated for this study were treated and followed at the Cullen Eye Institute in Houston, Texas, between 1970 and 1999. At the initial examination, demographic and other data were recorded onto medical record forms. The diameter (d) of the stromal infiltrate was generally measured with a slit-beam reticule or eyepiece micrometer. The infiltrate area was then estimated by πd2/4, rounding to the nearest 0.5 mm2. Additional information on risk factors, clinical features, laboratory data, interventions, and outcomes were collected from outpatient, hospital, photographic, microbiologic, and pathologic files and entered onto com- puterized spreadsheets. FIGURE 1A Climatic information was downloaded from the Case 15. Dematiaceous fungal growth on blood agar from corneal scrapings. online weather database provided by the National Climatic Data Center of the National Oceanic and Atmospheric Administration, US Department of Commerce. We averaged monthly data on temperatures recorded by 100 stations along the upper coast of Texas from 1970 through 1999. Corneal scrapings were routinely smeared onto glass slides for gram, Giemsa, acridine orange, and/or calcofluor white staining and were inoculated directly onto culture media that typically included a blood agar plate, a choco- late agar plate, a thiol or thioglycolate liquid, an anaerobic medium such as Brucella or Schaedler agar (each incubat- ed at 35°C), Sabouraud dextrose agar plate or slant, and brain-heart infusion (BHI) broth (both incubated at 25°C).89 The minimal requirements for laboratory confir- mation of Curvularia corneal infection were either a stained smear showing filamentous fungal elements with growth of Curvularia on at least 1 medium or isolation of Curvularia on at least 2 different primary media. Dematiaceous fungal growth was recognized as pig- mented colonies on C-streaks of primary culture media. FIGURE 1B Curvularia produced woolly olive-brown or black Case 15. Pigmented colonies on Sabouraud agar at several inoculation colonies, occasionally with a slate-blue sheen (Fig 1A and sites. 112 09 Wilhelmus Final 11/9/01 11:17 AM Page 113 Curvularia Keratitis trated fungal inoculum using a protocol approved by our institutional animal care and use committee. Ten NIH Swiss female mice (Harlan Sprague Dawley, Indianapolis, Indiana) were used, of which 4 were pretreated with intramuscular methylprednisolone (Depo-Medrol, Pharmacia & Upjohn, Kalamazoo, Michigan) 100 mg/kg 4 days before inoculation. Following anesthesia with keta- mine:xylazine:acepromazine, the cornea was scratched with a 25-gauge needle in a 6 x 6 linear grid pattern. A dark sludge of freshly grown C lunata, originally isolated from a human corneal infection (case 30), was prepared in phosphate-buffered saline to yield the spectrophotomet- ric equivalent of either 106 or 108 cells per 5 mL. One of these inocula was applied to the corneal surface, and the eyelids were rubbed together. Eyes were observed daily to detect corneal inflammation and were examined histopathologically after euthanasia. RESULTS FIGURE 1C Case 23. Septate hyphae and curved conidia on slide culture. CLINICAL SUMMARY or by a broth-dilution technique.93 The minimum fungi- Of 43 patients with Curvularia keratitis diagnosed by our cidal concentration (MFC) was determined by subcul- laboratory, 32 were treated at our institute (Table I). All tures from the MIC microwells onto blood agar plates. but 1 case began near the upper Texas or Louisiana coast Logistic regression assessed correlations between (Fig 2). Average patient age was 43 ± 21 years, including exposure variables